Keywords
Cervical cytology
Pap smear
Computer-aided diagnosis (CAD)
Cell classification
SIPaKMeD
How to Cite
Abstract
Manual review of cervical cytology specimens is labor-intensive and susceptible to fatigue-related errors, which has motivated the development of computer-aided diagnosis (CAD) systems for cervical cancer screening. To address the limited ability of conventional convolutional neural networks (CNNs) to capture global contextual relationships in cellular morphology, this study comparatively evaluates four contemporary Vision Transformer architectures, ViT-Base, DeiT-Base, PiT-Base, and Swin-Base, for five-class cervical cell classification. Experiments were conducted on the SIPaKMeD dataset comprising 4,049 isolated cell images. ImageNet-based transfer learning and strong regularization through RandAugment, Mixup, and CutMix were used to mitigate overfitting under limited-data conditions. Performance was assessed using accuracy, precision, recall, macro F1-score, parameter count, and computational complexity. ViT-Base achieved the best overall classification performance, with an accuracy of 96.90\% and a macro F1-score of 0.9691. PiT-Base provided the most favorable computational profile, requiring 72.74 million parameters and 21.1067 GFLOPs while maintaining an accuracy of 96.57\%. Error analysis showed that most residual confusion occurred between the morphologically similar metaplastic and koilocytotic classes. These findings indicate that Vision Transformer models can provide accurate and reproducible decision support for cervical cytology; however, external validation on multicenter and whole-slide data remains necessary before clinical deployment.
References
Altal, O. F., H. Gharaibeh, A. M. Sindiani, M. Amin, S. Alhatamleh, et al., 2026 Cervical cancer diagnosis using ct images: Enhancing the swin transformer model with global attention mechanism and stochastic depth. Biomedical Signal Processing and Control 116: 109519.
Chakraborty, G. S., M. Z. R. Saimon, M. Ahamad, P. Kondapalli, S. Das Polok, et al., 2025 An ensemble-based diagnostic framework integrating hybrid convolutional neural network models for automated cervical cancer detection: A multi-backbone comparative study and deep performance synthesis. pp. 27–32.
Deo, B. S., M. Pal, P. K. Panigrahi, and A. Pradhan, 2025 Cervigan: Cervical cancer pap smear image classification with gan-based data augmentation and external attention transformer. Discover Electronics 2: 73.
Devi, N. L., K. R. A. Britto, C. N. Savithri, and G. R. A. Jemila, 2026 Ncpls-vit: Enhancing vision transformer with nested patch scheduling for cervical cancer classification through contrastive self-supervised learning. Expert Systems with Applications 300: 130301.
Donepudi, R. and M. Kavitha, 2025 A hybrid cnn-bigru-gan framework for enhanced automated analysis of cervical cancer in medical imaging. International Journal of Advanced Computer Science and Applications 16: 1082.
Dosovitskiy, A., L. Beyer, A. Kolesnikov, D. Weissenborn, X. Zhai, et al., 2021 An image is worth 16x16 words: Transformers for image recognition at scale. In International Conference on Learning Representations.
Eric, S. M., B. Nassiri, Y. El Habouz, Y. El Mourabit, H. Tembine, et al., 2026 Lightweight hybrid mobilenet-vit architecture for efficient cervical cancer classification in cytological images. In Lecture Notes in Networks and Systems, volume 1852, pp. 217–230.
Hasan, M. N., S. B. Shuvo, M. M. H. Ankon, S. M. T. U. Raju, and N. Siddique, 2025 Transfusionnet: Framework for cervical cancer detection using deep learning with multi-level fusion. Results in Engineering 28: 107174.
Hossain, M. D., M. M. Rahman, M. A. Rahman, T. Ahmed, S. Dhar, et al., 2025 Cervitrans-xai: An explainable vision transformer ensemble for accurate cervical cancer classification. In 2025 IEEE 2nd International Conference on Computing, Applications and Systems.
Liu, Z., Y. Lin, Y. Cao, H. Hu, Y. Wei, et al., 2021 Swin transformer: Hierarchical vision transformer using shifted windows.
Mondal, J., R. Chatterjee, and M. K. Gourisaria, 2025 Vision transformer based approach for accurately detecting cervical cancer. In Proceedings of the 5th International Conference on Expert Clouds and Applications, pp. 500–506.
Mumuni, A., F. Mumuni, and N. K. Gerrar, 2024 A survey of synthetic data augmentation methods in machine vision. Machine Intelligence Research 21: 831–869.
Pradeep, J., M. Harikrishnan, P. Sanjay, R. Mohamed Siddiq, and N. Shivsharan, 2026 Multi-scale deep transfer learning and attention-guided feature fusion for robust cervical cell categorization. In Proceedings of the 2nd International Conference on Multi-Agent Systems for Collaborative Intelligence, pp. 656–662.
Ren, H., J. Guo, S. Cheng, and Y. Li, 2024 Pooling-based visual transformer with low complexity attention hashing for image retrieval. Expert Systems with Applications 241: 122745.
Sajiv, G. and N. Meenakshisundaram, 2025 Cervical cancer classification using transfer learning on pap smear images: A deep learning and explainable feature fusion framework. In International Conference on Communication, Computer and Information Technology.
Saju, S. and V. J. Panicker, 2026 A hybrid learning model for cervical cancer cell classification and detection. In Lecture Notes in Networks and Systems, volume 1670, pp. 159–171.
Shakil, M. R., A. H. Malik, M. I. H. Siddiqui, S. Ahmed, M. R. Miah, et al., 2026 Swin transformer-driven cervical cell classification with explainable ai and web-based screening. Journal of Medical and Health Studies 7: 25–35.
Sharma, N., K. Gaurav, and T. K. R. Bollu, 2025a Cervitransx: Explainable transformer-based cervical cancer classification. In Proceedings of the National Conference on Communications.
Sharma, N., K. Gaurav, and T. K. R. Bollu, 2025b Intrancenet: Interpretable and lightweight attention-based transformer architecture for cervical cancer diagnosis. In Proceedings of the International Joint Conference on Neural Networks.
Simaiya, S., U. K. Lilhore, Y. K. Sharma, L. P. Gorrepati, S. P. Kumar, et al., 2026 A hybrid deep learning framework for high-precision and efficient cervical cancer detection. IET Image Processing 20: e70300.
Sun, F., X. Ge, L. Deng, Y. Wang, H. Wu, et al., 2026 Research on a method for early diagnosis and progression prediction of cervical cancer based on imaging omics and molecular omics. Frontiers in Oncology 15: 1670852.
Tawfik, M., M. Heshmat, I. S. Fathi, G. Shakah, and W. M. Shaban, 2026 Federated learning-based cervical cancer classification using a novel hybrid kan-vit-autoencoder architecture.
Touvron, H., M. Cord, M. Douze, F. Massa, A. Sablayrolles, et al., 2021 Training data-efficient image transformers and distillation through attention. In Proceedings of the 38th International Conference on Machine Learning, volume 139.
Wubineh, B. Z., A. Rusiecki, and K. Halawa, 2025 Vit-se res: A hybrid vision transformer and resnet50v2 with squeeze-and-excitation block for cervical cell classification. In Lecture Notes in Computer Science, volume 15908, pp. 11–23.
Zhao, L., X. Shang, Z. Zhao, Y. Liu, Y. Liu, et al., 2026 Morphoformer: Dual-branch dilated transformer with pathological prior fusion for cervical cell morphology analysis. International Journal of Imaging Systems and Technology 36: e70273.
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.